A0805
Title: A bias-correction approach for cross-sectional and serial dependence in large panel model with interactive fixed effects
Authors: Runyu Dai - Tohoku University (Japan) [presenting]
Abstract: An efficient iterative principal components (IPC) estimator is developed for large linear panel models with interactive fixed effects. While conventional IPC estimators exhibit bias under cross-sectionally and serially correlated heteroskedastic errors, the proposed solution simultaneously addresses both issues through (1) a novel residual-based sparse regression approach for correlation correction and (2) standard heteroskedasticity adjustments. The estimator's asymptotic properties are rigorously established, and it is demonstrated through numerical studies its strong finite-sample performance in both estimation and inference.
